航空科学技术2025，Vol.36Issue(1)：33-38,6.DOI:10.19452/j.issn1007-5453.2025.01.004
基于卷积神经网络的呼吸变化趋势预测模型
Prediction Model of Respiratory Trend Based on Convolutional Neural Networks
摘要
Abstract
In the aerospace field,the physiological health of pilots is closely related to flight safety,and the comfort and oxygen utilization of breathing gas masks are key factors.Traditional aviation oxygen masks mainly use built-in sensors to complete lagging breathing gas valve adjustment,and users often feel that the mask has high breathing resistance and poor wearing comfort.In order to improve the pilot's operating experience and safety,this paper establishes a breathing change trend prediction model,aiming to provide a new solution for active regulation of breathing oxygen supply in respiratory protection.By proposing a prediction model based on convolutional neural network(CNN)algorithm,which predicts the outlet pressure data 15s later based on the corresponding breathing data collected by the inlet pressure and displacement sensors,the breathing change trend is predicted,thus achieving active oxygen supply and regulation of breathing.The research results show that the accuracy of the breathing change trend prediction model proposed in this paper is 92.422%.This model can improve the comfort and oxygen utilization of oxygen masks,and is of great significance for respiratory protection.关键词
呼吸预测/深度学习/卷积神经网络/主动调节Key words
respiratory prediction/deep learning/CNN/active regulation分类
信息技术与安全科学引用本文复制引用
李钰雯,田悦芃,戚晓玲,甘俊杰,黄栋,张志民..基于卷积神经网络的呼吸变化趋势预测模型[J].航空科学技术,2025,36(1):33-38,6.基金项目
航空科学基金(20200029069001) Aeronautical Science Foundation of China(20200029069001) （20200029069001）
